One account opens: the homepage, search, Marreta, the AI chat, Files, Gitea.
+One account opens: the homepage, search, Marreta, the AI chat, Files, Gitea, and Jellyfin +(there it is typed as username + password, not the automatic sign-in).
Three doors keep their own accounts and will not take the family -password: Jellyfin, Plex (a Plex account, its own e-mail -login) and the e-book door. Frank creates those by hand.
+Two doors keep their own accounts and will not take the family +password: Plex (a Plex account, its own e-mail +login) and the e-book door. Frank creates those by hand. Jellyfin does take the +family login — your franksdank.cloud username and password.
Want your own password? It lives at account settings — the same button you get at the bottom of the homepage.
@@ -128,8 +130,10 @@ account of any company's, and the posters and episode guides are all filled in.<Getting in
- Pick your name. There is no list of names on the login screen — type yours. -
- Type the password Frank set for you. It is not the family password; - this is the most common wrong door. +
- Type your franksdank.cloud password — the same one that opens the + homepage, search and the AI chat. There is no separate Jellyfin password. +
- Refused, but the same password opens the homepage? Your name hasn't been given library + access yet — ask Frank, it's a one-line change on his side.
- No name yet? Ask Frank to make one. There is no self-service signup, on purpose.
If it misbehaves
-
-
- Wrong password and no error you understand? That's the family password again. +
- It rejects a password that works everywhere else? Your name isn't on the library-access + list yet — ask Frank.
- A film with a blank poster is missing artwork, not missing files — Frank can refresh it.
- Buffering away from home usually means the pipe out of the house, not the server.
And the voices
+Two speech models live in this house, and they are different animals. One of them +speaks as a specific person; the other is a pack of stock voices that ships inside +the tools you already use.
+| Name | Runs on | Good at | Notes |
|---|---|---|---|
frank-tts |
+ the Strix Halo machine's built-in graphics, sharing the card with frank-dragoon |
+ cloning: it speaks as one particular voice, chosen by name | +five voices today — dr.kern, garrickhagon,
+ jeffhays, vaati, wilwheaton; ask for one by name and
+ you get it. No key, but it is a house-hardware door: it answers on the house network only |
+
| Kokoro | +your own browser, or Frank's own terminal — no server in the loop | +instant plain narration, quick and cheap | +about two dozen preset voices with coded names (af_heart,
+ bm_george…). It cannot clone, and it will never answer to the five names
+ above |
+
-
+
- The AI chat's speaker button is Kokoro. That is why it offers a + different list of names, and why it starts talking a moment before anything on the server + could. It speaks in your browser, on your machine. +
- The podcast maker uses the clones. An episode's two hosts are
+
wilwheatonandvaati, rendered byfrank-tts. That door + is Frank's for now.
+ - Frank's coding agent narrates through the clones too — the default
+ voice is
wilwheaton.
+ - Want your own voice in the list? Ask Frank. It needs about twenty + seconds of you talking plainly, once; the voice then carries your name. +
The exact weights: frank-tts is a Qwen3 talker
+(1.7 billion parameters, 4-bit) plus its 12 Hz audio codec, driven by
+qwentts.cpp; Kokoro is a much smaller 82-million-parameter model whose voices are
+baked in. A cloned voice is not a setting — it is a directory of reference audio on the
+server, which is why adding one is a fifteen-minute job and not a checkbox. Programmers on the
+house network: POST http://192.168.0.143:7772/v1/audio/speech with
+{"input": "…", "voice": "vaati"} gets you WAV or MP3, and
+/v1/audio/voices answers
+the live list, so this page can never lie to you for long.